David A. Bell
BSc, MPhil, PhD, DSc
A Little Bit...
David Bell is a computer scientist who is currently a Visiting Research Professor and Professor Emeritus at Queen’s University, Belfast (QUB). He has been a full professor for over 30 years. From 1986 to 2002, he was a Professor of Computing at Ulster University (UU), serving there as Head of the Department of Information and Software Engineering. At QUB he was Research Director of Knowledge and Data Engineering.
He has produced several hundred publications, and supervised around 40 PhDs to completion. He was prime investigator on many national projects and on many EU-funded projects (eg MAP, ESPRIT, DELTA, COST, AIM.) in IT in the Eighties and Nineties.
His research interests are centred on data and knowledge management - the linking of reasoning under uncertainty, machine learning, and other artificial intelligence techniques with database work . This involves the exploration of other topics in computing, including aspects of agent awareness and innateness.
You can read more here
Superintelligence and World Views
Current AI falls far short of producing 'better-than human' intelligence, which would arguably require genius-level ideas and world-views at least as rich as our own, including, for example, high quality evaluation and representation capabilities, in addition to very powerful hardware. They also lack other important things such as consciousness and even a sense of humour.
So humans stand apart from machines psychically as well as physically. The differences are primarily spiritual. Man-made superintelligence would challenge world-views which involve God and they would impact significantly on our human world-views in general.
Does all this rule out the possibility of their existence?
You can find more out about the book here
Purchase the book here
Main Contribution Areas
Complexity of placement
Methods for solving the Layered Data Placement Problem, include: the 'greedy', linear time, graph-collapsing method, the probabilistic hill-climbing method of Simulated Annealing, and a method based on Genetic Adaptation.
Hashing for data access
Probes/tuple = (hs + pg)/ps(h+g)
Formal Procedure Specification
Programming a computer is a useful vehicle for the study of this aspect of AI / problem-solving because the procedures can be automatically tested for correctness.
Distributed Database Systems
This book adopts a practical approach, reviewing the fundamentals of database technology and developments in data communications (including standards) before reviewing the principles of distributed DB systems. It includes case studies of the leading products.
Piecewise uniform distribution
Join size estimation
Concurrency Control in Distributed Databases and Distributed Expert Systems
Temporal and Image Systems management
Computational aspects of Dempster Shafer theory/ Rough Sets
Algorithm A has time complexity 0( IAI3 IUI2) to find one key
Pipelined Recursive query processing
The Boolean algebra of geometrical shapes can be used directly in Evidential Reasoning / Dempster Shafer theory(DS)
RԐS - Non-numeric Evidence Handling
Causality and Belief Networks
Combining Classifiers via DS
Machine Learning methods
KNN, SVM, Reinforcement, Neural Networks. etc
eg in Semantic Web
Get In Touch
Drop me an email
Email - firstname.lastname@example.org